Dental postoperative bleeding complications in ...

Oral Diseases (2012) doi:10.1111/j.1601-0825.2012.01922.x 2012 John Wiley & Sons A/S All rights reserved www.wiley.com

ORIGINAL ARTICLE

Dental postoperative bleeding complications in patients with suspected and documented liver disease CH Hong1, MW Scobey2, JJ Napenas1, MT Brennan1, PB Lockhart1 Departments of 1Oral Medicine, and 2Internal Medicine, Carolinas Medical Center, Charlotte, NC, USA

OBJECTIVES: The aims of this study were to determine the frequency of bleeding complications following dental procedures in patients with known or suspected chronic liver disease and whether international normalized ratio (INR) determination could aid in predicting bleeding complications in these patients. PATIENTS AND METHOD: We identified 90 patients (mean age: 51 ± 9 years) in this retrospective chart review. Sixty-nine patients had a known history of chronic liver disease and 21 had suspected chronic liver disease. Descriptive statistics were determined. Independent sample t-test and one-way variance test were utilized for continuous variables and chi-square test for dichotomous variables. RESULTS: The mean INR value for all patients was 1.2 ± 0.3. The INR value was significantly associated with the diagnosis of liver cirrhosis, the diagnoses of Hepatitis B and C together, the presence of ascites alone, and the number of clinical signs and symptoms (i.e. ascites, jaundice and encephalopathy) present. Nine patients with INR values between 1.5 and 2 underwent invasive dental procedures without postoperative bleeding complications. CONCLUSION: There were no episodes of postoperative bleeding in patients. The findings suggest that clinicians should not rely solely on an INR value to predict post-procedure bleeding in patients with liver disease. Oral Diseases (2012) doi: 10.1111/j.1601-0825.2012.01922.x Keywords: hemorrhage; hepatitis; liver cirrhosis; tooth extraction; dental prophylaxis

Correspondence: Catherine H Hong, BDS, MS, Orthodontics and Paediatric Dentistry, Faculty of Dentistry, National University of Singapore, 5 Lower Kent Ridge Road, Singapore 119074, Singapore. Tel: +65 6779 5555 ext. 1787, Fax: +65 6773 2602, E-mail: denchhl@ nus.edu.eg Received 3 January 2012; revised 26 January 2012; accepted 20 February 2012

Introduction A variety of diseases can lead to liver dysfunction and a concomitant coagulopathy (Friedman and Schiano, 2004; Golla et al, 2004; Firriolo, 2006; Roberts et al, 2010). Therefore, bleeding is a concern for dentists when patients with known or potential chronic liver disease require invasive procedures. It is essential that medical assessment of these patients include eliciting any history of liver disease, risk factors for chronic liver disease, or history of bleeding during and following surgical procedures (Lockhart et al, 2003a; Brennan et al, 2008a). A positive history for any of the above questions may dictate a review of the patient’s relevant laboratory values, or obtaining new studies. The combination of a thorough medical history as it pertains to coagulopathies, along with the appropriate laboratory values, should allow for an assessment of the risk for prolonged bleeding following an invasive dental procedure and the provision of appropriate preventive steps prior to, during, and following the procedure. There are various recommendations in the literature for the dental management of patients with liver disease, but they are largely based on expert opinion (Byron and Osborne, 2005; Firriolo, 2006). Furthermore, those recommendations do not address the full scope of the problem, to include patients with multiple coagulopathies. Most guidelines recommend that dentists assess the risk for prolonged post-procedure bleeding based on the severity of liver disease, the presence of other predisposing medical risk factors for bleeding (e.g. uremia), the use of anti-coagulant medications, and the specific dental procedure to be performed. Additionally, it is often important to obtain a complete blood count, platelet count, and coagulation parameters such as international normalized ratio (INR) as close as possible to the time of the invasive procedure. The INR is one of several common laboratory tests that can be utilized to assess coagulation. The INR has historically been used to monitor patients on warfarin, but it is also used to assess the coagulation status of patients with chronic liver disease. Depending on the reason for anticoagulation, the therapeutic INR

Bleeding in patient with liver disease CH Hong et al

2

range for patients on warfarin is between 2.0 and 3.5; while a normal INR is 1.0. The INR test result for patients taking warfarin should not be misinterpreted as having the same meaning as the INR value in patients with liver disease (Deitcher, 2002), for whom a small elevation in the INR value or prothrombin time implies significant liver disease (Lockhart et al, 2003a). This is likely because patients with liver disease have different and varying deficiencies in both vitamin K-dependent and vitamin K-independent procoagulant and anticoagulant factors compared to patients on warfarin therapy. What is also not clear is whether an elevated INR value predicts for the risk of bleeding from various dental procedures in this patient population. The primary goal of this study was to determine the frequency of bleeding complications after invasive dental procedures in patients with either documented or potentially undiagnosed liver disease. The secondary goal was to evaluate the utility of an in-office INR device to identify those patients at higher risk for bleeding.

Patients and methods Patients We retrospectively reviewed charts (electronic and paper) at our hospital-based outpatient dental clinic between 2004 and 2010 and identified 96 patients who had an in-office INR performed because of either suspected or documented liver disease. Six patients were excluded because of a concomitant secondary medical diagnosis that was a potential risk factor for bleeding. Two patients had acquired immunodeficiency disease syndrome, three had end stage renal disease (ESRD), and one was a kidney transplant recipient. The remaining 90 patients had a total of 108 dental appointments, of which 87 involved procedures with risk for postoperative bleeding. The mean age was 51 ± 9 years (range: 26–72 years), with 63 men (70%). Of the 90 patients, 69 (77%) had an in-office INR performed because of a documented history of chronic liver disease, 15 of whom were referred to the dental clinic by their physicians for dental evaluation prior to liver transplantation (Table 1). The remaining 21 (23%) had an INR performed because of their history of heavy alcohol consumption; of which two also presented with signs of jaundice. Overall, nineteen (21%) patients were still consuming large amounts of alcohol. Three patients (3%) reported a history of persistent bleeding after an invasive medical or dental procedure. Twenty-three (24%) were on medications that may potentiate bleeding: 10 were taking aspirin (aspirin 81 mg: 7; aspirin 325 mg: 3), two were taking clopidogrel (75 mg), three patients were on low molecular weight heparin (Lovenox, Sanofiaventis, Bridgewater, NJ, USA), and 10 patients were taking non-steroidal anti-inflammatory drugs (NSAIDs) on a regular basis. Two patients were taking more than one medication that could potentiate bleeding. Oral Diseases

Table 1 Documented Liver diagnosis in 69 patients Liver diagnosis Hepatitis B Hepatitis C Liver cirrhosis Alcohol Hepatitis C Alcohol and Hepatitis C Cryptogenic Autoimmune Hepatocellular cancer Others Drug-induced (acetaminophen) hepatitis Liver abscess Non-alcoholic steatohepatitis Primary sclerosing cholangitis Status after two liver transplants a

Patients Na 9 47 28 12 10 1 4 1 4 6 1 1 2 1 1

N > 69 as patients had more than one diagnosis.

Method Since early 2004, we have used an in-office INR test on all patients thought to be at risk for developing intra- or postoperative bleeding following invasive dental procedures. The inclusion criteria for this study included any of the following: (i) history of hepatitis B or C or other documented liver disease; (ii) abnormal liver laboratory values; (iii) history of prolonged bleeding after any invasive procedure; (iv) current or past history of heavy alcohol intake (>20 drinks per week for >2 years); and (v) clinical signs of jaundice and ⁄ or ascites and ⁄ or encephalopathy. The exclusion criteria included patients on warfarin therapy and those with concomitant secondary medical diagnoses that could potentially increase the risk for bleeding. We classified incision and drainage, extractions and hygiene procedures (e.g. scaling and root planning) as invasive dental procedures; these procedures carry a higher risk for postoperative bleeding compared to impressions, restorations, crown cementation, simple excisional biopsies, and removal of bone spicules. All INRs were performed using the CoaguChek System (Roche Diagnostics, Indianapolis, IN, USA) which has the advantage of providing point-of-care testing. There was a recall of the CoaguChek test strips by the United States Food and Drug Administration (FDA) in October 2006, during the course of our study. The FDA recommended duplicate testing for abnormal INR results, which we were already doing for quality assurance. No discrepancies in INR values were found based upon our duplicate testing. After identifying patients for inclusion, we extracted the following information from dental records, inpatient medical records, and electronic patient records for 25 hospitals and outpatient clinics in our hospital system: demographics, medical history, medications known to potentially interfere with platelet function (i.e. aspirin, clopidogrel, and NSAIDs), previous dental procedures performed, the use of local hemostatic agents and ⁄ or need for pre- and postoperative blood products at the time of previous dental procedures, and any reports of


postoperative bleeding. A postoperative bleeding complication was defined as any of the following: (i) visit to our dental clinic or one of our emergency departments with a complaint of postoperative oral bleeding within 2 weeks of the procedure; (ii) telephone call to our dental clinic with a complaint of postoperative bleeding within two weeks of the procedure, or (iii) documentation of bleeding at the postoperative visit. Descriptive statistics to include means, standard deviations, frequencies, and percentages were determined. Independent sample t-test and one-way variance (ANOVA) test were utilized for continuous variables and chi-square test for dichotomous variables. This study (#12-05-17E) was reviewed and approved by the Institutional Research Board at our hospital.

Results International normalized ratio values The average INR value for all patients was 1.2 ± 0.3 (range: 0.8–2.5). The mean INR value for the 69 patients with documented liver disease was 1.3 ± 0.3 (range: 0.9–2.5) (Table 1). Those referred for dental evaluation prior to liver transplantation (n = 15) had a mean INR of 1.5 ± 0.4 (range: 1.0–2.5). The mean INR value for the remaining 21 patients was 1.0 ± 0.1 (range: 0.8– 1.3). The INR value was significantly associated with the diagnosis of liver cirrhosis (P < 0.05) and diagnoses of both Hepatitis B and C together (P < 0.05). However, a lone diagnosis of either Hepatitis B (P = 0.414) or C (P = 0.07) and patient’s use of heavy alcohol abuse (past and ⁄ or present) (P = 0.07) were not associated with an elevated INR value. The INR value was significantly associated with ascites alone (P < 0.05) and the number of clinical signs and symptoms of ascites, jaundice, and encephalopathy present (P < 0.05). International normalized ratio values by procedure type Multiple invasive procedures were performed in eight appointments (Table 2). These included hygiene and dental restorations (n = 3), oral hygiene and tooth extractions (n = 2), oral hygiene and impression (n = 1), restorations and extractions (n = 1), and incision and drainage along with a single extraction (n = 1). Periodontal probing was only carried out in patients (n = 15) who were pre-liver transplant recipients. Dental treatment was deferred in seven and modified for three appointments for the following reasons. The mean INR value for these 10 appointments was 1.5 ± 0.4 (range: 1.0–2.0). Extractions were not performed because of high INR values of 1.7 and 2 in two deferred appointments. Dental treatment was deferred in five appointments as a result of patients’ request for sedation (n = 2), the need to clarify medical history with the physician, the absence of a legal guardian, and recruitment of the patient into a different research study within the department. In one of the modified appointments, extraction was deferred because of INR value of 2.0, but restorations were performed. In another mod-

3

Table 2 Dental procedures Type of dental procedures Examination (the status of periodontal health (i.e. probing depths) was not consistently evaluated) Restorations Hygiene Prophylaxis only Ultrasonic and hand scaling Scaling and root planning Extractions Single Multiple (range: 2–16 teeth) Local hemostatic measures Others Removal of bone spicule Biopsy Incision and drainage Impression Cementation of crown

Numbera 8 7 19 1 14 4 69 31 38 Sutures only: 11 Gelatin compressed sponge + sutures: 21 2 1 1 1 1

a Total number of procedures does not add up to 87 as multiple procedures were performed at some appointments.

ified appointment, full mouth gross debridement was not completed because of excessive bleeding (INR: 1.6). In the third modified appointment, fewer extractions were performed because of patient anxiety (INR: 1.5). Ten patients with INR values between 1.5 and 2.0 (mean: 1.7 ± 0.2) underwent invasive dental procedures. Of these 10 patients, platelet counts were available for four, the values range from 45 000 to 91 000 ll)1. Six patients received single or multiple extractions and all had sutures with or without the use of gelatin compressed sponges placed over the extractions sites. Of these six patients, one received both a single quadrant of scaling and root planning and a single tooth extraction. Four patients received hygiene procedures without any postoperative complications. Gross debridement, however, was discontinued in one patient mid-treatment (INR: 1.6) because of excessive perioperative bleeding (Table 3). Postoperative bleeding complications We did not find any records of patients who returned to the dental office or any of the hospital’s emergency or urgent care departments with complaint of postoperative bleeding complications.

Discussion The standard of care in our center has been to defer elective invasive dental procedures in patients with higher INR values (typically above 1.5) because of the level of suspected liver dysfunction. A discussion with their physician seems appropriate concerning the need for preoperative transfusion of blood products or additional tests for coagulopathies. The practice of deferment of treatment for patients with INR levels above 1.5 was extrapolated from a retrospective survey Oral Diseases


4

Table 3 Breakdown of international normalized ratio (INR) values by type of visits Visits with procedures with risk for bleeding

INR 1.5 (Gilmore et al, 1995). Many such guidelines, however, caution that elevated INR values in liver failure patients, while of prognostic significance for progression of disease, may not necessarily be a good indicator of bleeding risk (Segal and Dzik, 2005; Rockey et al, 2009). In our study, 10 patients with INR values between 1.5 and 2 underwent multiple extractions or subgingival scaling of teeth uneventfully and without the need for pre- or postoperative blood products. Given this observation and the overall absence of postoperative bleeding complications, our data suggest that uncomplicated extractions and hygiene procedures (e.g. dental prophylaxis, single quadrant scaling, and root planning) may be performed with the use of local hemostatic measures in liver disease patients with INR levels of up to 2.0, provided that there are no other contributing coagulopathies. For one patient, gross debridement had to be discontinued midway which suggests that some hygiene procedures, often thought to carry a lower risk of bleeding compared to extractions, should be thought of as highly invasive. This is a common misunderstanding but scaling, and especially root planning, involves a large surface area of potentially ulcerated mucosa by comparison with removal of a single tooth. Our decision to include patients on anti-platelet medications was based on a general consensus in the literature that these medications should not be stopped prior to minor surgical procedures in patients who are otherwise without risk factors for postoperative bleeding Oral Diseases

(Brennan et al, 2008b; Napenas et al, 2009). Of interest, we found no episodes of postoperative bleeding complications in this patient population who also had known or suspected chronic liver diseases. Of interest, two patients that were excluded from the study returned to the dental office with the chief complaint of persistent bleeding. One patient returned 2 days following two quadrants of scaling and root planning. The other patient, with an INR of 1.2 returned 13 days after extraction of seven teeth. His medical history was significant for hepatitis C, thrombocytopenia of unknown etiology (last documented platelet count a year prior was 79 000 ll)1) as well as ESRD requiring hemodialysis therapy. Both bleeding episodes, involved a mildly elevated INR with platelet count levels above 60 000 ll)1, which are considered to be sufficient for most dental procedures (Tripodi, 2006). It is noteworthy that an INR value of >1.2 was seen in several patients in our study who underwent various procedures but did not experience bleeding complications. This is in agreement with the study by Ward and Weideman (2006), which was the only published study we could find that evaluated similar parameters, but in pre-liver transplant patients. They reported that INR value was not statistically significant in predicting postoperative bleeding. However, both patients in our study with postoperative bleeding also had other risk factors (i.e. ESRD and use of clopidogrel) for platelet dysfunction that cannot be determined from the platelet count test (George and Shattil, 1991; Boccardo et al, 2004). This suggests that qualitative platelet dysfunction also plays a significant role in postoperative bleeding when combined with an


increased INR value and ⁄ or a low platelet count. Unfortunately, we were unable to evaluate the role of platelet function because of the retrospective nature of this study and the infrequency of bleeding complications. A systematic review by the British Committee for Standards in Haematology recommends that patients be questioned as to a history of bleeding because this can predict for the presence of a coagulopathy (Chee et al, 2008). Other authors have also suggested that a history of persistent bleeding or marked bruising may be just as important as traditional coagulopathy parameters in predicting the risk of post-procedure bleeding (Rockey et al, 2009). Both patients with bleeding complications who we excluded reported a positive history of persistent oral bleeding after a dental procedure. These findings suggest that INR values and quantitative platelet counts are of prognostic significance, but clinicians should not rely solely on an INR value or quantitative platelet count to predict post-procedure bleeding in liver disease patients with other diseases (e.g. renal or bone marrow) (Sreedhara et al, 1995; Lockhart et al, 2003a,b; Brennan et al, 2008b). From a review of the medical and dental literature, it is evident that the prediction and measurement of bleeding risk are difficult in patients with liver disease because of the lack of consistent correlation between conventional laboratory tests and hemostasis. Currently available guidelines, although lacking strong supportive evidence, often recommend platelet counts of at least 55 000 and 100 000 ll)1 for moderate risk and high-risk procedures, respectively. The measurement of fibrinogen levels and replacement if levels are